I am Sang Choe, a final year CS PhD student in
Language Technologies Institute at
Carnegie Mellon University, advised by
Eric Xing.
My research focuses on gradient-based optimization with the belief that gradient is a
key to understand various aspects of modern AI/ML, including generalization, uncertainty,
and interpretability. I also strive to understand gradient-based optimization from the
computer systems viewpoint, as, at the end of the day, what runs these AI/ML algorithms
are computers, requiring most practically useful algorithms to be highly compatible with
underlying systems. Last but not least, I aim to turn my research into highly scalable,
interoperable, easy-to-use softwares.
In technical terms, my research lies in the intersection of:
◆ machine learning debugging/auditing (data, weight, algorithm)
◆ automated machine learning
◆ machine learning software
Previously, I completed MS in Language Technologies at Carnegie Mellon University under the guidance of Jaime Carbonell. Before that, I earned BS in Electrical Computer Engineering & Mathematics (double major) from Seoul National University. I had also spent time as a research intern at Microsoft in 2021.
Making Scalable Meta Learning Practical
NeurIPS, 2023
Sang Keun Choe, Sanket Vaibhav Mehta, Hwijeen Ahn, Willie Neiswanger, Pengtao Xie, Emma Strubell, and Eric Xing
Betty: An Automatic Differentiation Library for Multilevel Optimization
[code]
ICLR, 2023
Sang Keun Choe, Willie Neiswanger, Pengtao Xie, and Eric Xing
Oral (1.8% acceptance rate)
Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning
[code]
OSDI, 2021
Aurick Qiao, Sang Keun Choe, Suhas Subramanya, Willie Neiswanger, Qirong Ho, Hao Zhang, Greg Ganger, Eric Xing
🏆 Jay Lepreau Best Paper Award
On Orthogonal Jacobian Regularization in Deep Neural Networks
SEDL Workshop @ NeurIPS, 2019
Sang Keun Choe*, Hosan Jeong*, Jaime Carbonell
On Leveraging the Visual Modality for Neural Machine Translation
INLG, 2019
Vikas Raunak*, Sang Keun Choe*, Quanyang Lu*, Yi Xu*, Florian Metze
Audio Cover Song Identification using Convolutional Neural Network ICASSP, 2017
Sungkyung Chang, Juheon Lee, Sang Keun Choe, Kyogu Lee